linux-kernel - Re: [PATCH] gpiolib: request fixes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Fri, 26 Dec 2008 07:53:14 -0800
From:	David Brownell <david-b@...bell.net>
To:	Magnus Damm <magnus.damm@...il.com>
Cc:	linux-kernel@...r.kernel.org, lethal@...ux-sh.org,
	akpm@...ux-foundation.org
Subject: Re: [PATCH] gpiolib: request fixes

On Thursday 25 December 2008, Magnus Damm wrote:
> From: Magnus Damm <damm@...l.co.jp>
> 
> Fix request related issues in gpiolib such as:
>  - fix request-already-requested handling in gpio_request()
>  - clear FLAG_REQUESTED on request error in gpio_direction_input()
>  - clear FLAG_REQUESTED on request error in gpio_direction_output()
> 
> Signed-off-by: Magnus Damm <damm@...l.co.jp>
> ---
> 
>  drivers/gpio/gpiolib.c |    3 +++
>  1 file changed, 3 insertions(+)
> 
> --- 0001/drivers/gpio/gpiolib.c
> +++ work/drivers/gpio/gpiolib.c	2008-12-26 13:09:50.000000000 +0900
> @@ -789,6 +789,7 @@ int gpio_request(unsigned gpio, const ch
>  	} else {
>  		status = -EBUSY;
>  		module_put(chip->owner);
> +		goto done;

Right, good catch.


>  	}
>  
>  	if (chip->request) {
> @@ -924,6 +925,7 @@ int gpio_direction_input(unsigned gpio)
>  			/* and it's not available to anyone else ...
>  			 * gpio_request() is the fully clean solution.
>  			 */
> +			clear_bit(FLAG_REQUESTED, &desc->flags);

NAK, this is insufficient ... it would need to drop the module
refcount and null the label too.  Plus this invalidates the
comment.  (Same below.)

However a basic premise is that drivers should now be avoiding
this legacy autorequest stuff, using gpio_request() instead.

Hence the comments here, below, and at ensure_requested() to
note the lack of cleanup if these legacy paths lose:  small
incentives to "do the right thing".  I'd rather see the work
go into making callers stop using autorequest; not making
that mechanism work better.

It may be time to make ensure_requested() use WARN(), which
will create a lot more noise than the current message ... a
larger incentive. :)


>  			goto lose;
>  		}
>  	}
> @@ -977,6 +979,7 @@ int gpio_direction_output(unsigned gpio,
>  			/* and it's not available to anyone else ...
>  			 * gpio_request() is the fully clean solution.
>  			 */
> +			clear_bit(FLAG_REQUESTED, &desc->flags);
>  			goto lose;
>  		}
>  	}
> 
> 


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/